Hello,<br>I am running a non-convex model. The information on the positive definiteness of the Hessian given by Ipopt does not indicate a local minimum. <br>I was wondering if running the model from a number of different starting points ("warm-start") will help to check the optimality of the solution. <br>
Given the fact that Ipopt is an interior point solver, will the "warm-start" option deteriorate its performance? <br>thanks in advance,<br clear="all"><br>-- <br>regards,<br>Hossein.<br>